AIExplained
patreon
“OpenAI is Not God”: DeepSeek, Liang Wenfeng and the R1 Phenomenon
🕑 Added 2025-04-03 22:27:01 +0000 UTCThe documentary revealing just how much the world got wrong about DeepSeek, what motivates the man behind the company, and what's next.
Incredible Editing: Hassan Iq.
Sources:
Liang Wenfeng 2023 Interview: https://web.archive.org/web/20241228030725/https://www.chinatalk.media/p/deepseek-from-hedge-fund-to-frontier
Liang Wenfeng 2024 Interview: https://web.archive.org/web/20241227111244/https://www.chinatalk.media/p/deepseek-ceo-interview-with-chinas
Hide from Crowds: https://www.theguardian.com/technology/2025/jan/28/who-is-behind-deepseek-and-how-did-it-achieve-its-ai-sputnik-moment
Fled the Province: https://www.ft.com/content/d23a01d8-4256-4142-a52c-a393ed6df1bc
DeepSeek V3 Paper: https://arxiv.org/pdf/2412.19437
DeepSeek R1 Paper: https://arxiv.org/pdf/2501.12948
Chat: https://chat.deepseek.com/
DeepSeek V1: https://arxiv.org/pdf/2401.02954
DeepSeek Coder: https://deepseekcoder.github.io/
DeepSeek Math: https://arxiv.org/pdf/2402.03300
DeepSeek MoE: https://arxiv.org/pdf/2401.06066
DeepSeek V2: https://arxiv.org/pdf/2405.04434
Anthropic CEO on DeepSeek: https://darioamodei.com/on-deepseek-and-export-controls?s=09
Liang Wenfeng: https://tv.cctv.com/2025/01/20/VIDEP7O5fkXufWLSm3J5pwkL250120.shtml
https://en.wikipedia.org/wiki/Liang_Wenfeng
Jim Simons: https://en.wikipedia.org/wiki/Jim_Simons
Altman Hopeless Quote: https://www.youtube.com/watch?v=AiE7FsdRzz8&t=3096s
1-2 two years behind quote: https://www.youtube.com/watch?v=1egAKCKPKCk
2-year Frenzy: https://www.wired.com/story/google-openai-gemini-chatgpt-artificial-intelligence/
Bing Sydney: https://www.nytimes.com/2023/02/16/technology/bing-chatbot-microsoft-chatgpt.html
Botched Bard: https://www.cnbc.com/2023/02/10/google-employees-slam-ceo-sundar-pichai-for-rushed-bard-announcement.html
Weights Visualized: https://hackaday.com/wp-content/uploads/2024/11/LLM-Anim-visualization.gif?w=800
MoE, credit Maarten Grootendorst: https://substackcdn.com/image/fetch/w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F101e8ddc-9aa7-4e24-92fc-78d25da73399_880x656.png
https://zilliz.com/blog/why-deepseek-v3-is-taking-the-ai-world-by-storm
Post Training credit Zhangjunlin: https://miro.medium.com/v2/resize:fit:1400/1*sCowcm-iJq9M3MOh6lorQg.png
Shoggoth Meme: https://huyenchip.com/assets/pics/rlhf/2-shoggoth.jpg
Value Model vs GRPO: https://community.aws/_next/image?url=https%3A%2F%2Fassets.community.aws%2Fa%2F2ra6RfOUylhUM8M3eDEMfDlB3l7%2FScre.webp%3FimgSize%3D1698x894&w=3840&q=75
Welch Labs: https://www.youtube.com/watch?v=0VLAoVGf_74&t=2s
Chip smuggling: https://www.theinformation.com/articles/nvidia-ai-chip-smuggling-to-china-becomes-an-industry?rc=sy0ihq
‘AI War’: https://scale.com/blog/win-the-ai-war
Doubao 1.5 Pro: https://www.aibase.com/tool/35837
Kimi 1.5: https://arxiv.org/pdf/2501.12599v1
o1-o3, OpenAI: https://openai.com/index/learning-to-reason-with-llms/
Semi-analysis on R1: https://semianalysis.com/2025/01/31/deepseek-debates/
Perplexity 1776: https://www.perplexity.ai/hub/blog/open-sourcing-r1-1776
OpenAI Counter-narrative: https://www.theguardian.com/technology/2025/jan/29/openai-chatgpt-deepseek-china-us-ai-models
DeepSeek Banned?: https://www.independent.co.uk/tech/deepseek-ai-us-ban-prison-b2692396.html
Guards to stop DeepSeek team leaving China: https://www.theinformation.com/articles/deepseek-national-treasure-china-now-closely-guarded?rc=sy0ihq
Comments
Kol Tregaskes
Great video, very well presented.
Antoine Ferrere
Youhoooooo can’t wait to watch! Thanks Philip!
Sunishchal Dev
The content of this video is obviously rock solid, as usual. This time, the soundtrack really stood out! Great work Philip, can't wait for the next one :)
